frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

FilesFlow – Browser-based PDF and image tools that run locally

https://filesflow.net/
1•blert•30s ago•0 comments

Why people chasing after useless token saving plugins and ignoring real solution

1•yohji1984•43s ago•0 comments

German AI consortium releases Soofi S, an open 30B model that tops benchmarks

https://the-decoder.com/german-ai-consortium-releases-soofi-s-an-open-30b-model-that-tops-benchma...
2•amai•2m ago•0 comments

The Hunt for the Counterfeiter Trying to Make the Perfect Bill

https://www.nytimes.com/2026/07/16/magazine/counterfeit-dollars.html
1•paulpauper•3m ago•0 comments

Funding Canonical Visions //GMs

https://amazing-lantern-c21.notion.site/Funding-canonical-visions-GMs-396adeecc3dc804ab012f015f3b...
1•paulpauper•3m ago•0 comments

The Decline of Sumptuousness in Cinema

https://yakimenko.substack.com/p/on-the-decline-of-sumptuousness-in-451
2•paulpauper•3m ago•0 comments

ACT-2 Preview: Sunday Robotics

https://twitter.com/tonyzzhao/status/2077806003308179802
1•manmeet•4m ago•0 comments

An overview of language features and implementation tasks that contribute toward

https://ghc.serokell.io/dh
2•fanf2•5m ago•0 comments

Teleprompter operator investigated over alleged Kalshi trades on Trump speeches

https://www.axios.com/2026/07/16/white-house-teleprompter-kalshi-investigation
1•doctoboggan•5m ago•0 comments

Show HN: Tree, truth, druid, dryad, and tar share one Proto-Indo-European root

https://p.migdal.pl/tree-of-tree/
1•stared•6m ago•0 comments

Flight Planning with Little Navmap

https://tech.marksblogg.com/little-navmap-flight-planning.html
1•marklit•6m ago•0 comments

Control of shear thickening rheological properties of polishing slurry(2021) [pdf]

https://www.euspen.eu/knowledge-base/ICE21153.pdf
1•pillars•6m ago•0 comments

Show HN: Custom Claude Status Lines

https://claudelines.com/
2•OG_BME•7m ago•1 comments

What can we learn from Bun's rapid Rust rewrite with AI?

https://newsletter.pragmaticengineer.com/p/the-pulse-what-can-we-learn-from-07f
1•pseudolus•7m ago•0 comments

It's shovels all the way down

https://jamesthomason.com/its-shovels-all-the-way-down/
2•dollar•8m ago•0 comments

FDA Approves First Oral Therapy That Inhibits PCSK9

https://www.fda.gov/drugs/news-events-human-drugs/fda-approves-first-oral-therapy-inhibits-propro...
1•michaefe•8m ago•0 comments

Cloudflare Outage Ongoing

https://slacreditwatch.com/i/cloudflare/kdnfshk5vs51
1•devd1976•9m ago•0 comments

Mushroom.app (Apple Intelligence, On-Device)

https://www.getmushroom.app/
1•qpe0•10m ago•0 comments

Visa Vulnerability Agentic Harness

https://github.com/visa/visa-vulnerability-agentic-harness
2•wslh•10m ago•0 comments

Guix: Creating a Package from a Binary

https://aloysberger.com/posts/guix-packaging-a-binary-as-a-guix-beginner.html
1•birdculture•11m ago•0 comments

Show HN: Dashi – Join the Social Analytics Experience

https://www.dashimetrics.com
1•launchpact_io•12m ago•0 comments

The Smith Chart: turning transmission line math into circles and arcs

https://www.nutsvolts.com/magazine/article/the-smith-chart
2•fcjr•13m ago•0 comments

Weston 16: HDR-ready, improved debugging, and DRM back end features

https://www.collabora.com/news-and-blog/news-and-events/weston-16-hdr-ready-improved-debugging-an...
2•losgehts•13m ago•0 comments

A Claude:// link could auto-submit hidden prompts in Claude Desktop

https://www.oasis.security/resources/reports/claude-url-scheme-prompt-injection
1•logickkk1•14m ago•0 comments

For Software Engineers, the AI Reckoning Is Here

https://www.bloomberg.com/news/features/2026-07-16/anthropic-and-openai-tools-transform-the-profe...
1•evo_9•15m ago•0 comments

Dallas Fed President calls for 'modestly' higher interest rates

https://www.cnbc.com/2026/07/16/dallas-fed-president-logan-calls-for-modestly-higher-interest-rat...
1•root-parent•15m ago•0 comments

I Made a Channel on Primal.net

https://primal.net/p/nprofile1qqsyj04902rxw3g9ecgjc9z0gtxa2pdsnuuh4ledv9n89gk29l23nus6au09l
1•doomsday1460•16m ago•2 comments

Show HN: HyperShots – Spec-perfect App Store screenshots from coding agents

https://github.com/hypersocialinc/hypershots
1•selcuk•17m ago•0 comments

Making Fable Cheaper Than Opus

https://twitter.com/joon_h_lee/status/2076714221837173097
1•gmays•17m ago•0 comments

Trump teleprompter operator allegedly made Kalshi bets

https://www.cnbc.com/2026/07/16/trump-kalshi-teleprompter-cftc-investigation.html
3•root-parent•17m ago•1 comments
Open in hackernews

Ask HN: Are we there yet? Can I create a human analogue using AI?

2•JoeAltmaier•1h ago
In Frederik Pohl's HeeChee series, a pivotal plot element was the idea of committing a human personality to an AI model. Are we there yet? What training data would be necessary or sufficient to cause an LLM to respond roughly as a human being? A particular human being, enough to pass something like the Is It Dad Turing test. There are potential products there, from hosting a memorial personality available to grieving relatives, to the re-personification of famous individuals. Yes, it is fairly morally ambiguous. But if there is money in it (agree or not) then it will be done.

Comments

JoeAltmaier•3m ago
I'd guess the training data would be a combination of baseline personality behaviors (turn-taking conversation, basic social skills, education level) with data specific to an individual. Recordings, letters, perhaps online posts but that's potentially not representative of the person since online is typically 'impersonal' by design.

Some of us may have too little to 'go on' to make a convincing avatar. Others may have biased samples, perhaps from lectures or recordings of performances.

Probably it may require an interview with those than knew them well. There would be memory bias, and the natural tendency to speak kindly of the dead. That might even serve the purpose better - it's a memorial after all, meant to be presentable and maybe not reflect the worst aspects of a personality.

gokuljs•2m ago
I might be wrong, but fine-tuning a model based on a particular human’s behavioural output should do the job to some extent, right?